[jcranmer]

Yeah, 32-bit x86 is somewhat easier because everything's in the same flat address space, and you at least have a system-wide code32 gdt entry that means you can ignore futzing around with the ldt. 16-bit means you get to deal with segmented memory, and the cherry on top is that gdb just stops being useful since it doesn't know anything about segmented memory (I don't think Linux even makes it possible to query the LDT of another process, even with ptrace, to be fair).

As for trying to ignore before main... well, the main benefit for me was being able to avoid emulating DOS interrupts entirely, between skipping the calls to set up various global variables, stubbing out some of the libc implementations, and manually marking in the emulator that code page X was 32-bit (something else that sends tools in a tizzy, a function switching from 16-bit to 32-bit mid-assembly code).

16-bit is weird and kinda fun to work with at times... but there's also a reason that progress on this is incredibly slow for me.

[evmar]

Slow progress is fine, it took me like two years to get where I got! (Not that I was working on it full time or anything, but also there were just many false starts and I had no idea what I was doing...)